Has traveledinternationallyto attend thisconferenceHas collaboratedon a researchpaper withsomeone from adifferent continentHas used agenerative AImodel for anon-academicpurposeIs optimisticabout thefuture ofhuman-AIcollaborationHas experiencewith low-resourcelanguages inNLPHas learneda newlanguage inthe last yearHascontributedto an open-source AIprojectHas presenteda paper onnaturallanguagegenerationHaspublishedresearch onmultilingualLLMsCan explain thedifferencebetween causaland maskedlanguagemodelsIs excitedabout thepotential ofLLMs ineducationHas apreferred AIresearch toolthey canrecommendKnows atleast threeprogramminglanguagesIs currentlyworking on aprojectinvolving cross-lingual transferlearningHas used agenerative AImodel tocreate art ormusicHasparticipated ina hackathonfocused on AIor LLMsIs familiarwith theconcept ofpromptengineeringCanrecommenda good AI ortech relatedpodcastHasexperiencewith fine-tuning a pre-trained LLMHasattended anICMLconferencebeforeHas used anLLM tosummarizeresearchpapersHassuccessfullydebugged acomplexLLMIs interestedin the ethicalimplicationsof generativeAICan namethreedifferent LLMarchitecturesHas traveledinternationallyto attend thisconferenceHas collaboratedon a researchpaper withsomeone from adifferent continentHas used agenerative AImodel for anon-academicpurposeIs optimisticabout thefuture ofhuman-AIcollaborationHas experiencewith low-resourcelanguages inNLPHas learneda newlanguage inthe last yearHascontributedto an open-source AIprojectHas presenteda paper onnaturallanguagegenerationHaspublishedresearch onmultilingualLLMsCan explain thedifferencebetween causaland maskedlanguagemodelsIs excitedabout thepotential ofLLMs ineducationHas apreferred AIresearch toolthey canrecommendKnows atleast threeprogramminglanguagesIs currentlyworking on aprojectinvolving cross-lingual transferlearningHas used agenerative AImodel tocreate art ormusicHasparticipated ina hackathonfocused on AIor LLMsIs familiarwith theconcept ofpromptengineeringCanrecommenda good AI ortech relatedpodcastHasexperiencewith fine-tuning a pre-trained LLMHasattended anICMLconferencebeforeHas used anLLM tosummarizeresearchpapersHassuccessfullydebugged acomplexLLMIs interestedin the ethicalimplicationsof generativeAICan namethreedifferent LLMarchitectures

Human BINGO: Navigating Generative AI and LLMs Across Languages - Call List

(Print) Use this randomly generated list as your call list when playing the game. There is no need to say the BINGO column name. Place some kind of mark (like an X, a checkmark, a dot, tally mark, etc) on each cell as you announce it, to keep track. You can also cut out each item, place them in a bag and pull words from the bag.


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
  1. Has traveled internationally to attend this conference
  2. Has collaborated on a research paper with someone from a different continent
  3. Has used a generative AI model for a non-academic purpose
  4. Is optimistic about the future of human-AI collaboration
  5. Has experience with low-resource languages in NLP
  6. Has learned a new language in the last year
  7. Has contributed to an open-source AI project
  8. Has presented a paper on natural language generation
  9. Has published research on multilingual LLMs
  10. Can explain the difference between causal and masked language models
  11. Is excited about the potential of LLMs in education
  12. Has a preferred AI research tool they can recommend
  13. Knows at least three programming languages
  14. Is currently working on a project involving cross-lingual transfer learning
  15. Has used a generative AI model to create art or music
  16. Has participated in a hackathon focused on AI or LLMs
  17. Is familiar with the concept of prompt engineering
  18. Can recommend a good AI or tech related podcast
  19. Has experience with fine-tuning a pre-trained LLM
  20. Has attended an ICML conference before
  21. Has used an LLM to summarize research papers
  22. Has successfully debugged a complex LLM
  23. Is interested in the ethical implications of generative AI
  24. Can name three different LLM architectures